重新认识 nginx 配置文件解析规则

需求

将除了

http://a.mengkang.net/read/(\d+)\.html
http://a.mengkang.net/simple/t(\d+)\.html           

之外的请求都转发到

https://mengkang.net

首页。

尝试解决

server  {
        listen  80;
        server_name  a.mengkang.net;
        root  /xxx;
        index index.html index.htm index.php;
    
        set $go2dev "1";
        if ($uri ~ "^/read/(\d+)\.html") {
                set $go2dev "0";
        }

        if ($uri ~ "^/simple/t(\d+)\.html") {
                set $go2dev "0";
        }

        if ($go2dev = "1") {
                rewrite ^(.*)$  https://mengkang.net? permanent;
        }

        rewrite ^/read/(\d+)\.html /read.php?tid=$1;
        rewrite ^/simple/t(\d+)\.html$ /simple/?t$1.html break;        
}           

测试发现

# 没问题
curl -LI http://a.mengkang.net/read/1.html 

# 跳转
curl -LI http://a.mengkang.net/simple/t1.html
HTTP/1.1 301 Moved Permanently
...
Location: https://mengkang.net           

很诡异第一个判断生效了，第二个判断没有生效么？

问题排查

if ($uri ~ "^/simple/t(\d+)\.html") {
-                set $go2dev "0";
+                rewrite ^(.*)$  https://baidu.com/?request_uri=$request_uri&uri=$uri? permanent;
        }           

再访问

http://a.mengkang.net/simple/t1.html

跳转到了

https://baidu.com

。说明规则没有问题。请求信息如下

curl -LI http://a.mengkang.net/simple/t14743.html
HTTP/1.1 301 Moved Permanently
Server: Tengine
Date: Tue, 15 Oct 2019 10:51:06 GMT
Content-Type: text/html
Content-Length: 278
Connection: keep-alive
Location: https://baidu.com/?request_uri=/simple/t14743.html&uri=/simple/t14743.html           

那问题到底出在哪呢？把刚刚的修改恢复，然后我在最后的跳转里增加一些日志，修改如下

if ($go2dev = "1") {
-                rewrite ^(.*)$  https://mengkang.net? permanent;
+                rewrite ^(.*)$  https://mengkang.net?request_uri=$request_uri&uri=$uri? permanent;
        }           

再次请求

curl -LI http://bbs.aliyun.com/simple/t14743.html
HTTP/1.1 301 Moved Permanently
Server: Tengine
Date: Tue, 15 Oct 2019 10:41:25 GMT
Content-Type: text/html
Content-Length: 278
Connection: close
Location: https://mengkang.net/?request_uri=/simple/t14743.html&uri=/simple/index.php

HTTP/1.1 200           

似乎发现了问题症结所在，第一次的请求确实没有进入

if ($go2dev = "1")

判断，直到执行了

rewrite ^/simple/t(\d+)\.html$ /simple/?t$1.html break;           

由重新匹配了一遍上面的规则，第二次匹配的时候，

$uri

发生了变化，不再是

/simple/t14743.html

而变成了

/simple/index.php

，所以进入最后的跳转规则里面。

只有把上面的规则做如下修改才不会导致循环匹配

- rewrite ^/simple/t(\d+)\.html$ /simple/?t$1.html break;
+ rewrite ^/simple/t(\d+)\.html$ /simple/index.php?t$1.html break;